4 Applications of the Multispectral Mixing Scheme
4.1 Case 1: Low power LEDs cluster design
We setup a multispectral platform composed of four single-colour LEDs (Excellence Opto. Inc., EOQ-5ERF red, EOQ-5EYF amber, EOQ-5DFE green, EOQ-5EBF blue) and a phosphor-converted cool-white LED (Excellence Opto. Inc., EOQ-5EWF). The corresponded spectra at ambient temperature Ta of 25 oC and drive currents IPWM of 20 mA are shown in Figure 4-1. An adequate layout of LED arrangement and optics with first-order design are considered to deliver a uniform illumination on the measured plane . Due to the relative low level of drive currents (maximum 20mA) controlled by the pulse-width modulation (PWM) approach, the modeling of the spectral power distribution SPD for each colour LED can be assumed to satisfy the scalability and additive property in color mixing scheme , which indicates the SPD can be thermal -independent and be regularly proportional to drive currents without distortion.
Figure 4-1 The spectra of red (R), green (G), blue (B), amber (A) and cool-white (CW) LEDs at ambient temperature Ta of 25 oC with all drive currents of 20 mA. The corresponded chromaticity points and specifications are also shown in the figure. The drive currents controlled by PWM approach have the pulse width of 6.66 ms at differences of 0.04 ~ 0.06 ms for each gray level (a total of 128 gray levels).
4.1.1 Validation of the composite spectrum
As the aforementioned process, we start to fulfill the models of multispectral LEDs through optimization step with weight factors w = 0 (efficiency mode) and 1 (quality mode). The resultant spectra for color temperatures CT = 3000K and 6500K are selected to be verified as shown in Figure 4-2, where the simulations are in close agreement with the experimental measurements by checking the R2 value (above 0.98) and the chromaticity deviation xy (below 0.01). The Figure 4-3 features illuminant environments for both color temperatures 3000K and 6500K at the efficiency mode.
Apparently the appearance of the color checker chart is more accurate in the booth with higher CQS than that in the booth with poor CQS (refer to the difference of the grey scale coloured squares between these booths).
Figure 4-2 Spectral comparisons of simulations and experiments for P0 and P1 at CTs of 6500K and 3000K, respectively. The simulated spectra closely
Figure 4-3 The illuminant environments at (a) P0 (CQS = 87 points, LE = 66 lm/watt) for CT= 6500K and (b) P0 (CQS = 69 points, LE = 67 lm/watt) for CT = 3000K show apparently different color rendering abilities.
4.1.2 Comparison of R/G/B and R/G/B/A system
In addition to the experimental validation, more insight can be pursued by the quantitative analyses. Here we assume the minimum requirements for color rendering CQSm = 80 points and luminous efficiency LEm = 60 lm/watt. Based on the linear approximation SA2 method, the loci of R/G/B, R/G/B/A, and R/G/B/A/CW are presented in Figure 4-4. The black curve in Figure 4-4(a) depicts the referenced single solution of R/G/B cluster for color temperature CT from 1000K to 10000K. It can be noted that all of the solutions are far from the required performance (quadrant I). We than add the amber (A) source to the R/G/B cluster, the average improvement of 50%
in color quality scale CQS is achieved without too much loss of luminous efficiency (LE1/0 < 5%). This result is generally in accordance with the concept that the color rendering performance of sources would be improved when its modulated spectrum is as smooth as sunlight.
For the solution of each color temperature CT, the point P0 (w = 0) is set as the starting point as we mentioned in the Section of merit analysis. In the view point of P0, the information of the increment rate of CQS denoted as CQS1/0 and the sacrifice of
decrement rate of LE denoted as LE1/0 can be given in Figure 4-5. The result of the R/G/B/A case shows that all circle points for whole range of color temperature CT are located at right-top corner. Thus the designer would undoubtedly chose a high weight value (w ~ 1) to boost the color rendering ability with a little expense of cluster efficiency. This tendency is equivalent to drive P0 to approach P1 along the straight line in Figure 4-4 (a). Nevertheless, the R/G/B/A cluster still suffers a stringent operating window of CT from 2800K to 3000K, which would strictly preclude its use in intelligent lighting applications.
Figure 4-4 The SA2 results of (a) R/G/B (black curve), R/G/B/A, and (b) R/G/B/A/CW clusters aimed to P1 and P0 for full range of CT from 1000K to 10000K.
Figure 4-5 The results of CQS1/0 and LE1/0 for R/G/B/A and R/G/B/A/CW clusters. By using SA2 analysis, R/G/B/A/CW can further extend the operation window in color temperature.
4.1.3 The effect of cool-white LED
Compared with the R/G/B/A cluster, the addition of cold-white (CW) LED is able to further extend the operational window throughout the entire color temperature range.
To prove this statement we can select a start point P0 at CT= 3000K, whose CQS0 = 66.8 points (unqualified) and LE0 = 66.7 lm/watt [refer to Figure 4-4 (b)]. The corresponded information of CQS1/0 = 34.3% and LE1/0 = −2.4% at the same point P0
can be found in Figure 4-5. With above parameters, it is readily to derive an appropriate weight of 0.79 by using Equation (3-17) to fulfill the requirement, where the CQS value is increased to 85 points at the expense of 1.3 lm/watt.
Generally, the weighting value can be conducted to the comparison between CQS1/0 and LE1/0. That means the balance condition of CQS1/0 ≈ −LE1/0 at color temperature CT of 5200K (Figure 4-5) could be regard as a turning point for the weight selection. By applying this weight value selection strategy to the R/G/B/A/CW combination, it is logically to choose a large weight value (w > 0.5) for CT < 5200K
and vice versa in order to approach the requirements. In sum, adapting the phosphor- converted white source could further increase 5% in CQS and 20% in LE over full range of color temperature, which is due to the cool-white LED associated with high efficiency offers a good candidate to substitute the function of blue colour. The detail will be analyzed as follows the illustration of Figure 4-6.
Figure 4-6 (a) The values of CQS and LE, and (b) the stacked emission power ratio versus color temperature for the optimized R/G/B/A/CW design (CQSm = 80 points and LEm = 60 lm/watt). The operation window has been extended to 2600K < CT < 8500K with the selected weight via SA2 selection method. It is noted that the operation window is mainly restricted by the CQS due to the correction factor at the extreme color temperature.
4.1.4 The color tunable R/G/B/A/CW system
At this point, we can successfully determine the operation point by the proposed methodology and set an optimal lighting environment for R/G/B/A/CW system. As shown in Figure 4-6, the operation window is extended to span across 2600k − 8500K with the user-defined requirements of CQSm = 80 points and LEm = 60 lm/watt, which would be shrink by more severe lighting requirement accordingly (e.g. the operation window of 3200 − 5600K for CQSm = 90 points and LEm = 64lm/watt).
Based on Figure 4-6, we can find that when the color temperature is less than the 6400K, the power ratio is mainly governed by the light quality requirement, and each component has a comparable amount. On the other hand, the efficiency requirement is dominated and contributed by cool-white LED when the operation temperature is higher than 6400K. The combination of LED cluster reduces to R/G/CW for 6400K <
CT < 8500K as shown in Figure 4-6 (b). Within the operation window of 2600K ― 8500K, the function of blue LED has been replaced by the cool-white light source, so that we can discard it from the cluster for most general lighting applications.